In This Issue
Jason Droege, Benchmark Partner, former founder and head of Uber Eats. He joined Uber in 2014, starting from scratch to explore new businesses within the platform, and within 6 years built Uber Eats from zero to a global GMV run rate of $20 billion. Droege's core judgment is: the key to building a "second curve" within a platform is not perfect institutional design, but the CEO's "blind faith" in the person in charge — the willingness to bear the consequences of possible failure, which is the prerequisite for entrepreneurs to dare to take risks.
Theme 1: The Key to Intrapreneurship within a Platform — CEO's "Blind Faith" Rather than Exquisite Incentive Design
Droege argues that the true prerequisite for successful intrapreneurship within a large company is the CEO's trust in the person in charge, not a perfect incentive contract.
- Historical context: At the end of 2013, Uber's management realized that the "local logistics" dimension held even greater possibilities, but the specifics were unclear. CEO Travis Kalanick believed there was a "big thing" and also believed Droege could find it, so he gave him a blank slate. Droege described this as "willing to bear the consequences of your failure" — he once burned through $30 million during the exploratory phase with little to show for it, but the CEO defended him at the board level.
- Mechanism breakdown: Droege believes it is impossible to precisely replicate entrepreneurial incentives with a paper contract. The problem is that tying incentives to specific KPIs such as revenue or geographic expansion would strip the person in charge of the flexibility to make the right decisions. "What you need to know is that the CEO will take care of you — if this works out, you will get more money than you need. This trust should exist before all conversations."
- Deduction: Droege contrasts this model with the reality of most companies — quarterly reviews, annual bonuses, internal processes all drain entrepreneurial momentum. Companies that can provide this kind of "entrepreneurial license" need to have enough stock or value to support it. Falsification condition: If the CEO is not prepared to defend a massive failure (such as burning tens of millions of dollars) in front of the board, this model should not be attempted.
Theme 2: Lessons from Competitive Boundaries — A 180-Degree Turn from "Speed First" to "Supply First"
Droege believes that the biggest mistake in Uber Eats' early days was being dominated by Uber's "cultural bias" — prioritizing speed above all else, while ignoring what customers truly wanted: supply choice.
- Historical Context: Uber's corporate culture was "the faster, the better" — "If we can get you a car in 5 minutes, we can get you anything in 5 minutes." This belief led to the first product, Uber Fresh: placing pre-made food in cars, delivering meals like a ride-hailing service. It turned out that although lunch-time efficiency was extremely high (drivers made 5-6 deliveries per hour), business growth hit a ceiling.
- Data Chain: Customers repeatedly said "we want more restaurants, better restaurants," but the team kept asking "how much faster can we go?" It wasn't until late 2015 when they launched Eats in Toronto — with 50-100 restaurants to choose from — that the first 3 hours of sales exceeded the total of the previous 4 days (of the Uber Fresh product).
- Mechanism Breakdown: Droege distilled the causal chain of "belief → thinking → priority → action → result." If you don't examine your own biases, you will go further and further in the wrong direction. He later systematized this lesson: before every major decision, first ask "why do we think this way? What do we believe?"
- Competitive Landscape: Droege observed that competitors DoorDash and Postmates adopted a "list first, sign later" strategy early on — first listing all the most popular restaurants, then seeing which had demand before signing them. Caviar, on the other hand, focused on high-end restaurants. Droege believes that "supply constraints" are the best signal — those restaurants that are most popular but have limited supply (single or double locations) are the core drivers of platform growth.
- Deduction: Droege believes that if the right product (choice first) had been launched from the start, Uber Eats' market share could have been higher. Falsification Condition: If a market still cannot improve unit economics or customer retention after supply is sufficient, it indicates that the market itself may have issues.
Theme 3: Competition and Exit — Lessons from the Indian Market: The Dual Thresholds of Speed and Capital
Droege believes that in markets with strong network effects, first-mover advantage is extremely important, and the lesson from the Indian market is that even outstanding team execution cannot compensate for a 10-month delay and a disadvantage in capital.
- Data chain: Droege originally planned to enter 45 countries within 24 months, but because he felt it was "too many," he delayed entering the Indian market by about 10 months. When he decided to postpone, Swiggy had only raised around $14 million; by the time he decided to enter, Swiggy had received $50–100 million from Naspers ("unlimited capital"). Zomato had also secured deep-pocketed backers.
- Mechanism breakdown: Competition in the Indian market evolved into a "discount war every 30 minutes" — three competitors (Swiggy, Zomato, Uber Eats) optimized dynamic discounts at the finest granularity to fight for a single basis point of market share. In Droege's view, the Indian team were "absolute beasts" as competitors, but "even if you execute better, you cannot outspend your opponent."
- Simulation and exit: Ultimately, Uber sold its India business to Zomato. Droege considers this a "correct but reluctant" exit. Falsification condition: If, in a market, an opponent's fundraising ability and capital endurance far exceed your limits, even a better product and execution cannot win — this is the signal to exit.
Theme 4: Core Insight of Platform Economics — Relative Leverage and Pricing Philosophy
Droege believes that successful marketplace platform founders must understand 'relative leverage' — that is, the price sensitivity of both supply and demand sides, and how to dynamically adjust profit allocation among the three sides (consumers, restaurants, drivers).
- Mechanism breakdown: Droege's pricing strategy in Toronto was to 'start with a price higher than necessary and see if the market accepts it' — charging restaurants a 30% commission and consumers a $5 delivery fee. The result proved that the market could accept it. He believes that unit economics 'will take care of themselves' because consumer demand is strong enough, and the incremental demand for restaurants (incrementality) is sufficient to support a high commission rate (eventually stabilizing at 24-25%).
- Data chain: By mid-2017, 19 months after launch, Uber Eats had reached $2 billion in GMV and was 'barely losing money.' Droege believes this is evidence that the business itself was already healthy before the competitive and funding environments changed.
- Competitive landscape: Droege analogizes the concept of 'relative leverage' to 'financial/stock market dynamics' — you need to know that 'when I charge consumers a little more, I can charge restaurants a little less; when I charge restaurants less, more restaurants will join.' He looks for founders who can 'dive into the details' and understand this balance, rather than just having a grand vision of 'connecting supply and demand.'
- Deduction: Falsification condition: If, after market expansion, unit economics (such as restaurant commission rate, consumer delivery fee, and driver efficiency) cannot converge to a sustainable level, it indicates that the theoretical 'relative leverage' hypothesis may not hold. Droege's confidence is based on early tests (such as the change from free to a $5 fee in Toronto within 3 weeks, observing the change in retention rate).
Theme 5: A Unique Perspective on Failure – Distinguishing 'Good Failure' from 'Bad Excuses'
Droege disagrees with the Silicon Valley popular narrative that 'failure is good.' He believes most failures are avoidable, and that failure becoming a 'cultural meme' instead encourages inefficiency.
- Mechanism Breakdown: Droege categorizes failure into three types:
1. Business opportunity exists but execution is poor – This is bad and unacceptable.
2. Business opportunity exists but science/technology is not yet mature – This is the most forgivable.
3. The business opportunity simply does not exist, but you mistakenly believe it does.
- He argues that in the consumer internet and platform space, 95% of ideas are technically feasible, so failure is usually not a 'cutting-edge science problem' but a 'judgment and execution problem.'
- Extrapolation and Risk: Droege warns that in an environment of excess capital, the saying 'failure is good' becomes 'an excuse to justify failure.' Falsification conditions: If the team can demonstrate after failure 'why this failure was inevitable' (e.g., technical limitations, external force majeure), he is willing to accept it; but if it is merely 'we tried but didn't succeed,' he tends to see it as an execution problem.
- Extension – Ideas vs. Execution: Droege believes that 'ideas are important, but execution is 95%.' He disagrees with the view that 'ideas are worthless and can be casually shared' – he keeps the best ideas confidential and only shares them with clients. Because 'if you have a truly good idea, you want to have as much execution time window as possible.'
Mentioned Companies
| Company |
Guest View (Positive/Risk Warning/Neutral) |
Key Data |
| Uber Eats |
As a founder, highly positive about its business model |
Mid-2017 GMV approximately $2 billion, almost no loss; the author estimates without COVID it could have reached $3 billion |
| DoorDash |
As a competitor, neutral – recognizes execution strength |
Used the "list first, sign later" strategy to acquire supply, comparing its strength |
| Zomato |
Indian competitor, cautious about its competitive intensity |
Collaborated with big backers like Naspers, Uber eventually exited the Indian market (sold to Zomato) |
| Swiggy |
Indian competitor, cautious about its competitive intensity |
Funding from $14 million (when Droege decided to delay) to $50 million–$100 million (when entering) |
| Caviar |
Early competitor, neutral |
Focused on high-end restaurants, eventually acquired by DoorDash |
| Postmates |
Early competitor, neutral |
Entered the market before Uber Eats, later merged with Uber Eats |
| Instacart |
Neutral – appreciative |
As a "case of successful collaboration with large suppliers," Droege said "they did it, incredible" |
| McDonald's |
Neutral – appreciative |
World's largest restaurant chain, partnered with Uber Eats in 2017, Droege said it was "rarely earlier than most peers" |
Judgments Worth Remembering
1. “You must risk the end of your career to find the success that defines your career” — Jason Droege
- Supporting evidence: This is not a slogan, but his practice at Uber — the CEO defended him at the board level for burning $30 million on a failure. Falsification condition: If the CEO does not have the authority or willingness to make such a defense, this model does not hold.
2. “Uber believes in you more than you believe in yourself” — Jason Droege
- Supporting evidence: This is central to Uber’s early culture: letting employees (especially those on the front lines) dare to make decisions, because the cost of a decision is relatively small, while the delay (opportunity cost) is huge. He gives an example: If Uber Eats had launched a year later, GMV might have shrunk from $50 billion to $20 billion.
3. “Speed first is a cultural bias, not a market truth” — Jason Droege
- Supporting evidence: Uber believed that “delivering anything in 5 minutes” was most important, but customers consistently demanded “more restaurants.” The Toronto Eats product (100 restaurants) had 3-hour sales equal to the previous 4 days’ sales. Falsification condition: If in a market, after supply is sufficient, unit economics and retention rates do not improve, it indicates that supply is not the problem, but price or experience.
4. “In India, even if you execute better, you cannot outspend your competitors” — Jason Droege
- Supporting evidence: Swiggy went from a $14 million funding round to $50 million–$100 million (with Naspers entering), and the Indian market became a “discount war every 30 minutes.” Falsification condition: If the competitor’s funding capacity is something you cannot catch up to, the signal to exit is the competitor’s “unlimited capital.”
5. “Success is not determined by a single disruption, but by continuously optimizing supply-side matching and marketing strategy” — Jason Droege
- Supporting evidence: He summarizes that “supply constraints + popularity” are the key to success, not “the most well-known brand.” He cites the “after hours” demand (hot dog shops after 11 PM) as an example.
6. “Failure should not be celebrated, but should be highly scrutinized” — Jason Droege
- Supporting evidence: He categorizes failures into three types, arguing that 95% of consumer internet ideas are technically feasible, and failure is usually an execution problem rather than a science problem. The name of this framework: Droege’s “Failure Taxonomy” — “Business opportunity exists but execution is poor” (unacceptable), “Business opportunity exists but technology is not mature” (forgivable), “Business opportunity does not exist” (misjudgment).
7. “Uber Eats was almost not losing money before mid-2017” — Jason Droege
- Supporting evidence: This was achieved within 19 months, with GMV of approximately $2 billion. He uses this to refute the narrative that “food delivery platforms can never be profitable,” arguing that “hyper-growth brings hyper-pain, but the pain will be digested at some point.”
8. “80% of the global restaurant market is fragmented” — Jason Droege
- Supporting evidence: Even McDonald’s (which accounts for 1% of the global population’s daily visits) has only a single-digit market share in the US. The platform’s value lies in aggregating fragmented supply. Falsification condition: If the platform cannot establish sufficient “relative leverage” among fragmented suppliers, it cannot form a moat.